Re: [PATCH 2/2] locking/percpu-rwsem: Introduce bias knob

[Oleg Nesterov]

On 07/14, Peter Zijlstra wrote:
>
> +void rcu_sync_sabotage(struct rcu_sync *rsp)
> +{
> +	rsp->gp_count++;
> +	rsp->gp_state = !GP_IDLE;
> +}

Ah, I didn't notice this !GP_IDLE...

Please use GP_PASSED, this is what this actually means. And note the
wait_event(GP_PASSED) in rcu_sync_enter().

Otherwise

Reviewed-by: Oleg Nesterov <oleg@xxxxxxxxxx>